106 EXPERIMENT STATION RECORD. 



official (the Liudo Gladding- and the alternate) and Stassfiirtor Anlialt 

 methods are pointed out. An error common to all is due to the 

 tendency of the precipitates (particularly barium sulphate) thrown 

 down in the course of the method to occlude a certain amount of potasli. 

 In addition to this, in the Lindo-Cxladdiug method the alcohol washings 

 remove some of the potash, \yith the use of the ammonium-chlorid 

 solution it is claimed that a double decomposition results by which 

 the potash of the potassium-platinic chlorid is replaced by ammonia. 

 Moreover the same change occurs when the ammonium-chlorid solu- 

 tion is saturated with potassium-platinic chlorid. By following the 

 official directions an ammonium-chlorid solution was obtained which 

 contained ''about as much potash in the form of chlorid as platinic- 

 chlorid.'' The author's observations indicate, moreover, that the 

 potassium-platinic chlorid obtained by this method is never strictly 

 pure. "In addition to ammonium-jilatinic chlorid, calcium and mag- 

 nesium sulphate, and sometimes other impurities, are almost certain 

 to "be present. . . . The alternate method is to a certain extent liable 

 to the same objection as the Liudo-Gladding. Most of the loss here 

 comes from the occlusion of potash in the barium sulphate." 



The German method is believed to require less manipulation, and 

 after facility in its use is acquired is considered preferable to the other 

 methods. The author is "convinced, however, that it is chargeable 

 with the same essential defects as the Lindo-Gladding and the alter- 

 nate methods." 



Tables are given which show the amounts of potash recovered from 

 the alcohol and ammonium chlorid washings in the Lindo-Gladding 

 method and from the ammonia and ammonium oxalate and barium 

 sulphate precipitates in the alternate method. Tlie average loss to 

 each 1 per cent of potasli was found to be 0.029 per cent. 



The quantitative determination of carbohydrates, E. Schulze 

 {Chem. Zt(f., 18 {1894), No. 29, pp. 527, 558).— Although cane sugar, 

 milk sugar, aud starch can be quantitatively determined by inversion 

 with dilute mineral acids and titration with Fehling's solution, it is 

 explained that this method does not give exact results except when 

 all the prescribed conditions are carefully observed, and presents 

 serious difficulties when a mixture of carbohydrates is present, for the 

 following reasons: 



(1) Tlie inversion does not progress with Cipial rapidity witli dif- 

 ferent sugars; even between carbohydrates, which are otherwise simi- 

 lar, there are considerable differences. 



(2) The inversion is often accomiianied by a cliange known as 

 "reversion," through which complex compounds are produced, the 

 majority resembling disaccharids or dextrins. 



(3) By continued heating with dilute H.2SO4 or HCl the glucoses 

 are gradually decomposed, forming humin substances aud certain acids 

 (levulinic, formic, etc.). 



